I’m excited to share a project I’ve been working on called AudioMuse-AI, an open-source containers app that brings smart playlist / similar-song discovery to your media server setup—especially if you’re running Jellyfin or Navidrome.
The GitHub repository is this:
https://github.com/NeptuneHub/AudioMuse-AI
Audiomuse-AI in short:
- Uses Librosa + TensorFlow to do sonic analysis — extracting features like tempo, energy, moods, embeddings, etc. 
- From that data, it can generate:
- similar-song searches
- “song path” playlists between two tracks
- “sonic fingerprint” playlists based on your listening habits 
- It’s containerized (Docker / Podman) and also deployable via Kubernetes (Helm chart provided) for folks who like to run their own infrastructure. 
This is currently in beta (just passed from alpha). This means you can be one of the early adopter or maybe contribute in its success.
The vision of AudioMuse-AI is bring Sonic Analysis free and opensource to everyone, without need of payment of any kind. If you like just drop a star ⭐ on the repo and maybe raise an issue with your feedback!
Thanks everyone!
Top comments (1)
Hi, I recently found your distilled clap model and ended up integrating it into a mobile offline music player for radio queues of individual songs, albums, and artists. It runs really really well on my smartphone. Just wanted to pop in and say kudos for your hard work!